AQDavid(22)
7/22/2005
8:34 AM
Page 216
PROVING CAUSATION WHERE THE
BUT FOR TEST IS UNWORKABLE
Hillel David, W. Paul McCague, Peter F. Yaniszewski*
1. Introduction
On a practical level, causation1 simply means that the current
condition or circumstances would be different had an act or omission
not occurred.2 The alteration in circumstances can be positive, negative,
or just a maintenance of the status quo. The critical matter is that the
situation would not be what it is had there been no act or omission;
otherwise, the act or omission cannot be said to have had any effect on
the current situation. The “but for” test is merely another way of
expressing this concept of change or difference in the current situation
that would not otherwise have been present. The same concept may be
expressed in the statement that the act or omission was a necessary
factor, or that the injury was inevitable.
The rationale for the critical importance of the but for test is
contained in the elegantly simple remarks of Major J.:3
[O]ne need only consider first principles. The essential purpose and most
basic principle of tort law is that the plaintiff must be placed in the position he or she would have been in absent the defendant’s negligence (the
“original position”). However, the plaintiff is not to be placed in a position better than his or her original one. It is therefore necessary not only
to determine the plaintiff’s position after the tort but also to assess what
the “original position” would have been. It is the difference between
these positions, the “original position” and the “injured position”, which
is the plaintiff’s loss.
*
1.
2.
3.
Of McCague, Peacock, Borlack, McInnis & Lloyd LLP.
The term “causation” in this article will be a reference only to causation-in-fact,
and not to the secondary causation issue involving remoteness in law.
The reference to the “occurrence” of an omission is shorthand for the defendant’s
failure to perform the act(s) that he or she ought to have performed.
Athey v. Leonati (1996), 140 D.L.R. (4th) 235 at para. 32, [1996] 3 S.C.R. 458,
[1997] 1 W.W.R. 97 (original emphasis).
216
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 217
Proving Causation where But For Unworkable 217
In the same decision, however, the statement was made that in those
instances where the but for test is unworkable, “causation is established where the defendant’s negligence materially contributed to the
occurrence of the injury”.4 There are two problematic aspects to that
statement which will be addressed below. In addition, this article poses
the question: Is it time for a further relaxation of the rules regarding the
manner of proving causation?
2. What Is Causation and Why Is It a Necessary Element?
The dictionary meaning of “cause” is “a person or thing that gives
rise to an action, phenomenon, or condition”.5 That definition is
reflected in the need for the injury to be attributable to the defendant’s
wrongdoing6 or, as otherwise stated, the need to prove that “the defendant’s impugned conduct actually caused the loss complained of”.7
Reference has also been made to the concept of “making a difference”
to the outcome.8 In an extra-judicial writing, Sopinka J. said:9
The concept of causation is in many respects the foundation of liability in
tort law. It is the link between the conduct of the defendant and the plaintiff’s loss. This connection has frequently been expressed primarily in
terms of a physical cause and effect relationship. As society and tort law
evolve, it is apparent that causation is better understood as the expression
of that degree of connection between the defendant’s conduct and the
plaintiff’s loss which is sufficient to warrant making the defendant pay.
It is a basic principle that “A defendant in an action in negligence is not
a wrongdoer at large: he is a wrongdoer only in respect of the damage
which he actually causes to the plaintiff.”10 A similar remark is that
4.
5.
6.
Athey, ibid., at para. 15.
New Oxford English Dictionary (New York, Oxford U.P., 2001).
Allen (Next Friend of) v. University Hospitals Board, [2002] A.J. No. 1041 (QL)
at para. 38, [2003] 1 W.W.R. 621, 281 W.A.C. 59 sub nom. Allen v. University
Hospitals Board (C.A.).
7. Stewart v. Pettie, [1995] 1 S.C.R. 131 at para. 60, 121 D.L.R. (4th) 222, [1995] 3
W.W.R. 1 (emphasis added). See also paras. 65 and 69.
8. Chisholm v. Liberty Mutual Group (2002), 60 O.R. (3d) 776 at para. 25, 217
D.L.R. (4th) 145, 163 O.A.C. 129 (C.A.); Loder v. Somerton, [2003] N.J. No. 16
(QL) at para. 53, 221 Nfld. & P.E.I.R. 199 (S.C.), affd [2004] N.J. No. 79 (QL),
235 Nfld. & P.E.I.R. 42 (C.A.); Cameron (Litigation Guardian of) v. Louden,
[2000] O.J. No. 858 (QL) at paras. 315 and 423 (S.C.).
9. The Honourable John Sopinka, “Whither Causation” (Insight Educational
Services, June 18, 1991). In Loder, ibid., Orsborn J. said at para. 58: “Given the
finding of negligence, it is the causal relationship that justifies the payment of
compensation.”
10. B.M. v. British Columbia (Attorney General), [2004] B.C.J. No. 1506 (QL)
at para. 157 (C.A.), quoting from Browning v. War Office, [1962] 3 All E.R. 1089
(C.A.).
AQDavid(22)
7/22/2005
8:34 AM
Page 218
218 The Advocates’ Quarterly
[Vol. 30
“one is never simply liable; one is always liable for something”.11
Causation is intended to establish the “substantial connection”,12 the
“sufficient link”13 between the wrongdoing and the injury.
The but for test of causation “. . . serves as an exclusionary test. Its
purpose is to eliminate from consideration factually irrelevant causes . . .
If the but for test is not met then the injury would have occurred
regardless of the act or omission in question.”14 It would be illogical to
require a person, in a fault-based system, to pay compensation for an
injury that would have occurred in any event. Causation is therefore an
essential element in a tort cause of action for damages.15
3. The But For Test
“The general, but not conclusive, test for causation is the ‘but for’
test, which requires the plaintiff to show that the injury would not
have occurred but for the negligence of the defendant.”16 It is a
“traditional principle in the law of torts”17 and applies equally in the
law of contract, at least insofar as non-disclosure is concerned.18
A “but for” cause has also been referred to as a “proximate” cause.19
It “has been traditionally applicable even in cases where the
hypothetical question requires prediction of human reaction”.20
11. Fairchild v. Glenhaven Funeral Services Ltd., [2003] 1 A.C. 32 at paras. 12 and
54 (H.L.) (original emphasis).
12. Snell v. Farrell (1990), 72 D.L.R. (4th) 289 at p. 299, [1990] 2 S.C.R. 311, 107
N.B.R. (2d) 94; Hock (Next Friend of ) v. Hospital for Sick Children, [1998] O.J.
No. 336 (QL) at para. 116, 106 O.A.C. 321 sub nom. Hock v. Hospital for Sick
Children (C.A.).
13. Lurtz v. Duchesne, [2003] O.J. No. 1540 (QL) at para. 164, 122 A.C.W.S. (3d)
384 (S.C.) (varied on other grounds 136 A.C.W.S. (3d) 1055 (Ont. C.A.), released
February 4, 2005), echoing the remarks of Sopinka J. quoted above.
14. Chisholm, supra, footnote 8, at para. 25.
15. Although there may be an exception in the case of fraud — see Parallels
Restaurant Ltd. v. Yeung’s Enterprises Ltd. (1990), 4 C.C.L.T. (2d) 59 at para. 40,
49 B.L.R. 237 (B.C.C.A.), application to review granted 26 A.C.W.S. (3d) 223
where evidence of the inevitability of the loss was held to be irrelevant. This may
have been an example of the comment: “Fraud unravels everything.” Sylvan Lake
Golf & Tennis Club Ltd. v. Performance Industries Ltd. (2002), 209 D.L.R. (4th)
318 at para. 68, [2002] 1 S.C.R. 678, [2002] 5 W.W.R. 193.
16. Athey, supra, footnote 3, at para. 14.
17. Snell, supra, footnote 12, at pp. 293-94.
18. Hodgkinson v. Simms (1994), 117 D.L.R. (4th) 161 at pp. 203 and 209, [1994] 3
S.C.R. 377, 57 C.P.R. (3d) 1.
19. Blaz v. Dickinson, [1996] O.J. No. 3397 (QL) at para. 76, 23 M.V.R. (3d) 70
(Gen. Div.).
20. B.M. v British Columbia (Attorney General), supra, footnote 10, at para. 177,
although those are among the types of situations where the “but for” test may well
be unworkable: Walker Estate v. York-Finch General Hospital (2001), 198 D.L.R.
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 219
Proving Causation where But For Unworkable 219
The term “. . . does not mean ‘exclusive’. It means, in effect,
‘necessary’.”21
In some circumstances, and particularly where there are multiple
independent causes,22 the but for test is unworkable23 — in other words,
it is practically impossible for the plaintiff to prove by the usual means
that the same injury would not have occurred had the defendant’s
wrongdoing not taken place. Another way of putting it is that it may be
practically impossible for the plaintiff to show that he or she would
have been in a better position had there been no wrongdoing. One
example of practical impossibility of proof is the question whether a
person’s marriage would have failed regardless of the defendant’s
sexual abuse of that person.24
The question then arises: How is the plaintiff to prove the necessary
element of causation when the circumstances are such that the but for
test, satisfaction of which is a necessary ingredient of causation, is
unworkable? In the landmark Athey decision, that question was
answered with the problematic statement referred to in the introduction
above. It is reproduced here in full:25
The “but for” test is unworkable in some circumstances, so the courts
have recognized that causation is established where the defendant’s
negligence “materially contributed” to the occurrence of the injury.
The first concern with that statement is that it implies that the but
for test is to be abandoned in cases where it is unworkable and in its
place is to be substituted the “materially contributing cause” test. That
response has been repeatedly adopted in the cases since Athey.26
21.
22.
23.
24.
25.
26.
(4th) 193 at para. 88, [2001] 1 S.C.R. 647 sub nom. Walker Estate v. York Finch
General Hospital, 145 O.A.C. 302.
Phillip v. Whitecourt General Hospital, [2004] A.J. No. 1342 (QL) at para. 247,
359 A.R. 259 (Q.B.).
Walker Estate, supra, footnote 20, at para. 87.
Athey, supra, footnote 3, at para. 15.
B. (E.) v. Order of the Oblates of Mary Immaculate in the Province of British
Columbia, [2001] B.C.J. No. 2700 (QL) at para. 235, 110 A.C.W.S. (3d) 289 (S.C.),
revd on other grounds 227 D.L.R. (4th) 298, [2003] 7 W.W.R. 421, 300 W.A.C. 288
(C.A.).
Athey, supra, footnote 3, at para. 15.
Among many other cases, see: H.L. v. Canada (Attorney General), [2005] S.C.J. No.
24 (QL) at para. 123; Walker Estate, supra, footnote 20, at para. 88; B.M. v British
Columbia (Attorney General), supra, footnote 10, at para. 124; Briglio v. Faulkner,
[1999] B.C.J. No. 1293 (QL) at paras. 33-35, 204 W.A.C. 92, 69 B.C.L.R. (3d) 122
(C.A.); MacCabe v Westlock Roman Catholic Separate School District No. 110,
[2001] A.J. No. 1278 (QL) at para. 52, [2002] 1 W.W.R. 610, 257 W.A.C. 41 sub
nom. MacCabe v. Board of Education of Westlock (Roman Catholic Separate)
School District No. 110 (C.A.); Mizzi v. Hopkins (2003), 64 O.R. (3d) 365 at para.
31, 171 O.A.C. 161, 37 M.V.R. (4th) 50 (C.A.); Lurtz v. Duchesne, supra, footnote
AQDavid(22)
7/22/2005
8:34 AM
Page 220
220 The Advocates’ Quarterly
[Vol. 30
That cannot, however, have been intended, because any type of cause,
whether it be the cause (a sole cause) or a cause (one of multiple causes),
incorporates by definition the “but for” concept. Otherwise, it would not
be a cause in any normal sense of that term. To abandon the but for test
is to abandon the element of causation. There is nothing in the decision
that suggests an elimination of the need for the element of causation,
including its embedded “but for” component. To the contrary, the “first
principles” comment27 made in the same decision unambiguously establishes the requirement for satisfaction of the but for test.
Situations in which the but for test is unworkable are those in which
causation cannot be proved at the scientific level. The law, however,
does not require that level of proof. The issue is not whether, in those
circumstances, the but for test is irrelevant or is not a necessary ingredient of causation. The issue in such situations rather involves the
manner and standard of proof that will satisfy, for legal as opposed to
scientific purposes, the necessary element of causation, including its
core component as expressed by the but for test.
The following statement too indicates that the fact that the but for
test is unworkable in a given situation does not call for the abandonment of the element of causation in that situation:28
The question that this court must decide is whether the traditional
approach to causation is no longer satisfactory in that plaintiffs in
malpractice cases are being deprived of compensation because they
cannot prove causation where in fact it exists.
The concept expressed by the but for test is a plain and unavoidably
necessary component in showing that the wrongdoing “contributed”, in
the sense of being a contributing cause, to the loss or injury. This leads
to the second difficulty with the statement quoted above. It appears to
amount to little more than the comment that “causation is established
when causation (in the form of a materially contributing cause) is established”. A factor must first be shown to have been a cause before it can
be said to have been a materially contributing cause.
13, at para. 347, vard Ont. C.A., February 4, 2005; Hartwick v. Simser, [2004] O.J.
No. 4315 (QL) at para. 214 (S.C.); Gill Estate v Marriott, [1999] O.J. No. 4509 (QL)
at para. 123, 93 A.C.W.S. (3d) 376 (S.C.); Chesher v. Monaghan, [1999] O.J. No.
1278 (QL) at para. 54, 87 A.C.W.S. (3d) 824 (Gen. Div.); Hampton v. Marshall,
[1997] B.C.J. No. 301 (QL) at para. 43, 68 A.C.W.S. (3d) 1066 (S.C.), supp. reasons
70 A.C.W.S. (3d) 906; Phillip v. Whitecourt General Hospital, supra, footnote 21, at
paras. 244-52; Whitfield v. Calhoun, [1999] A.J. No. 357 (QL) at para. 88, 242 A.R.
201 (Q.B.); Hetu v. Traff, [1999] A.J. No. 714 (QL) (Q.B.) at para. 34; Fournier v.
Wiens, [2004] A.J. No. 643 (QL) at para. 96, 33 Alta. L.R. (4th) 114 (Q.B.).
27. See supra, footnote 3.
28. Snell, supra, footnote 12, at p. 298 (emphasis added).
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 221
Proving Causation where But For Unworkable 221
The statement is unfortunately a misplaced response to a different
question (in fact, two different questions) than the one initially posed
by the court. It is relevant not to the issue of how to deal with situations in which the but for test is unworkable, but rather to the different
issues of: (a) whether one out of multiple causes will satisfy the need
for the element of causation and, if so, (b) what the strength of the
causal connection or impact must be.
The response to the quandary created by situations in which the but for
test is unworkable is actually provided later in the Athey reasons. That
response involves a relaxation of the rules regarding the manner, but not
the standard, of proof of causation. By not eliminating the need for the
element of causation, Athey did not, implicitly at least, do away with the
but for test. Satisfaction of that test is still necessary for the purpose of
establishing the element of causation. It is in regard to the manner of
establishing that element, and of the but for component in particular, in
situations in which proof by usual and normal means may be practically
impossible, that Athey is a seminal decision. The but for test remains, in
all tort claims, a component of the indispensable element of causation.
Subject to exceptions that arise on policy grounds,29 liability should
be imposed only where causation, whether in the form of a sole cause
or a materially contributing cause, in fact exists. Causation can exist
only where the but for test has been satisfied. This means that the but
for test should not be abandoned, but rather that the means by which
the test can be satisfied should be relaxed so that, for example, a factor can be found to be a “but for” (or necessary) sole or contributing
cause on the basis of an inference drawn as a matter of common sense.
The court in Athey approved a relaxation of the rules regarding the
manner of satisfying the test for causation.
The correct approach (apart from further confusion caused by a
different meaning given to the term “material”) is outlined in the
following comments:30
Where a plaintiff proves that the defendant’s breach of duty materially
contributed to the injury, that is sufficient to prove that that the breach
caused the injury.
.....
29. As will be discussed later, the only such exceptions at present are (a) the situation
where the same injury would have occurred in any event because of another
person’s wrongdoing, that wrongdoing not being an unrelated intervening cause,
and perhaps (b) cases involving fraud (see footnote 15).
30. Couillard v. Waschulewski Estate (1988), 59 Alta. L.R. (2d) 62 at p. 77, [1988] 4
W.W.R. 642, 87 A.R. 161 (Q.B.), followed in Gallant v. Fialkov (1989), 69 O.R.
(2d) 297 at p. 316, 50 C.C.L.T. 159 (H.C.).
AQDavid(22)
7/22/2005
8:34 AM
Page 222
222 The Advocates’ Quarterly
[Vol. 30
If proof by a plaintiff that the defendant’s breach of duty materially contributed to his injury relieves the plaintiff of his obligation to show that
the defendant’s breach caused the injury, then it must follow that a
material contribution to an injury is one that but for such contribution the
injury would not have occurred. By saying what a material contribution
does, i.e., satisfies the plaintiff’s ordinary proof of causation, Wilsher is
confirming the “but for” test for causation. It is because “but for the
contribution the injury would not have occurred” that the contribution is
designated “material”.
The result of the misplaced, and therefore non-responsive, response in
Athey has been that courts have posed the question whether, in cases
where the but for test is unworkable, there was a materially contributing
cause, rather than asking the correct question, that being whether
causation, including its embedded component, the but for factor, has
been established by way of some relaxed means of proof. They have,
in other words, been asking the wrong question.
For the most part, they have also failed to ask the question: Is it necessary, in the circumstances of the particular case, to expand even further
— i.e. beyond the parameters laid down in Athey — the relaxation of the
means by which the but for test can be taken to have been satisfied?
4. Cases Involving a Failure to Act
Both acts and omissions can constitute negligent conduct. The “but
for” articulation is ill suited for cases where the negligence involves
omission rather than commission. In such instances, the question
“Would the same injury not have been sustained but for the defendant’s
negligence?” will inevitably be answered in the affirmative — “Yes, it
would have been sustained regardless of the defendant’s failure to act”
— because the defendant’s negligence consists of the failure to prevent
or alleviate an injury that is otherwise bound to occur. The appropriate
formulation in such cases is whether the same injury would have
occurred “even if” the defendant had acted.
The following comment has been made regarding causation in the
context of negligent omissions:31
In a practical sense, it is much more difficult to determine what would
have happened if something which ought to have been done but was not
done had been done than it is to determine the effect of the doing of
something which ought not to have been done.
31. Miller v. Budzinski, [2004] B.C.J. No. 2718 (QL) at para. 471, 136 A.C.W.S. (3d)
1141 (S.C.), quoting from Lankenau v. Dutton (1991), 79 D.L.R. (4th) 705,
[1991] 5 W.W.R. 71, 55 B.C.L.R. (2d) 218 (B.C.C.A.), leave to appeal to S.C.C.
refused 83 D.L.R. (4th) vii.
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 223
Proving Causation where But For Unworkable 223
Some cases in which causation has been found where the defendant’s
negligence involved a failure to act are listed below.32
5. Multiple Causes and Materially Contributing Causes
It is not now necessary, nor has it ever been, for the plaintiff to establish
that the defendant’s negligence was the sole cause of the injury . . . The
law does not excuse a defendant from liability merely because other
causal factors for which he is not responsible also helped produce the
harm.33
A factor is a contributing cause for legal purposes if it “materially
contributed” to the occurrence of the injury, and a contributing factor is
“material” if it falls outside the de minimis range,34 although other cases
have made reference to a “substantial” connection or contribution.35 The
de minimis range has not been defined and its meaning clearly depends
on the circumstances of each case. Some guidance may be obtained
from decisions that equate that term with a matter that is not trivial or
insignificant,36 with the term “insignificant” itself meaning “that which
is of no importance; trivial; trifling; contemptible”.37
There is no requirement that a materially contributing cause be a
sufficient38 cause,39 although some cases indicate, although not with
32. Kamloops (City) v. Nielsen (1984), 10 D.L.R. (4th) 641, [1984] 2 S.C.R. 2, [1984]
5 W.W.R. 1; Stewart v. Pettie, supra, footnote 7; B.M. v British Columbia
(Attorney General), supra, footnote 10; Meloche v. Hotel Dieu Grace
Hospital/Villa Marie, [1999] O.J. No. 3525 (QL), 179 D.L.R. (4th) 77, 124 O.A.C.
330 (C.A.); Swanson Estate v. Canada (1991), 80 D.L.R. (4th) 741, [1992] 1 F.C.
408, 7 C.C.L.T. (2d) 186 (F.C.A.); Doe v Metropolitan Toronto (Municipality)
Commissioners of Police (1998), 39 O.R. (3d) 487, 160 D.L.R. (4th) 697, 126
C.C.C. (3d) 12, sub nom. Jane Doe v. Metropolitan Toronto (Municipality)
Commissioners of Police (Gen. Div.); Fournier v. Wiens, supra, footnote 26; Scott
v. Mohan, [1993] A.J. No. 592 (QL), 142 A.R. 281, 142 A.R. 281 (Q.B.).
33. Athey, supra, footnote 3, at paras. 17 and 19. See also Derksen v. 539938 Ontario
Ltd., [2001] 3 S.C.R. 398 at paras. 30-32, 205 D.L.R. (4th) 1, 153 O.A.C. 310;
Mizzi v. Hopkins, supra, footnote 26, at para. 28.
34. Athey, ibid., at para. 15.
35. Swanson Estate v. Canada, supra, footnote 32, at p. 757; Roncato v. Caverly
(1991), 5 O.R. (3d) 714 at pp. 717, 720, 723 and 724, 84 D.L.R. (4th) 303, 53
O.A.C. 367 (C.A.); Blaz v. Dickinson, supra, footnote 19, at para. 81; Lindahl
Estate v. Olsen, [2004] A.J. No. 967 (QL) at paras. 123 and 128 (Q.B.).
36. R. v. Nette, [2001] 3 S.C.R. 488 at paras. 1-14, 205 D.L.R. (4th) 613, 158 C.C.C.
(3d) 48; R. v. Malmo-Levine; R. v. Caine, [2003] 3 S.C.R. 571 at paras. 113-34,
233 D.L.R. (4th) 415, 179 C.C.C. (3d) 417.
37. R. v. Nette, ibid., at para. 7.
38. Meaning a cause that would have led to the same injury even had there been no
other contributing causes.
39. Athey, supra, footnote 3, at para. 17; Myers v. Peel County Board of Education
(1981), 123 D.L.R. (3d) 1 at p. 12, [1981] 2 S.C.R. 21, 17 C.C.L.T. 269; Reeves v.
AQDavid(22)
7/22/2005
8:34 AM
Page 224
224 The Advocates’ Quarterly
[Vol. 30
any great clarity, that the attribute of sufficiency is required.40 It is
difficult to imagine how a factor that would qualify as a materially
contributing cause in that it barely exceeds the de minimis range could
also be a sufficient factor, so that to require that attribute would be
inconsistent with the de minimis test.
Complicating matters further is the question whether a materially
contributing cause must satisfy the “but for” test. As earlier indicated,
it is our view that a factor cannot be a cause in any sense of that term
unless it satisfies the but for test (although how to prove such satisfaction in situations where the test is unworkable on the scientific level is
a different issue, and one which will be dealt with below). There are
cases that hold, at least impliedly, that a materially contributing cause
can only be one that does satisfy that test,41 although one Law Lord in
the recent Fairchild decision has taken the opposite view.42
Subject to what is said below regarding a sufficient non-tortious
cause and contributory negligence, the finding of a materially contributing cause makes the defendant liable for the full extent of the
plaintiff’s injury, whether the other contributing causes were tortious
or non-tortious, so long as it is not a divisible injury, which would
include a part of the injury arising from an unrelated intervening cause.
The only difference between the two situations (tortious vs. nontortious contributing causes) is that apportionment among tortfeasors
is available under legislation such as, for example, the Ontario
Negligence Act.43 Where the plaintiff has been contributorily negligent
40.
41.
42.
43.
Arsenault, [1998] P.E.I.J. No. 95 (QL) at para. 19, 168 Nfld. & P.E.I.R. 251
(C.A.), leave to appeal to S.C.C. refused 193 Nfld. & P.E.I.R. 182n; Hetu v. Traff,
supra, footnote 26, at para. 36; Dushynski v. Rumsey, [2001] A.J. No. 792 (QL) at
para. 147, [2001] 9 W.W.R. 327, 94 Alta. L.R. (3d) 26 (Q.B.), vard [2003] 9
W.W.R. 72, 296 W.A.C. 373, 16 Alta. L.R. (4th) 237.
Walker Estate v. York-Finch General Hospital, supra, footnote 20, at para. 88; X
v. R.D.M., [2004] B.C.J. No. 2044 (QL) at para. 144 (S.C.); B. (E.) v Order of the
Oblates of Mary Immaculate, supra, footnote 24, at para. 221 (S.C.); Fairchild v.
Glenhaven Funeral Services Ltd., supra, footnote 11, at para. 170.
B.M. v British Columbia (Attorney General), supra, footnote 10, at para. 187;
Roncato v. Caverly, supra, footnote 35, at pp. 718-20; Swanson Estate v. Canada,
supra, footnote 32, at pp. 757 and 760; Gallant v. Fialkov, supra, footnote 30, at
p. 316; Wilson v. Byrne, [2004] O.J. No. 2360 (QL), 131 A.C.W.S. (3d) 962 (S.C.)
at para. 82; Phillip v. Whitecourt General Hospital, supra, footnote 21, at para.
248, although the contrary view was expressed at paras. 249 and 252.
Fairchild v. Glenhaven Funeral Services Ltd., supra, footnote 11, Lord Rodger at
para. 129.
G. (E.D.) v. Hammer, [2003] 2 S.C.R. 459 at paras. 29-32, 230 D.L.R. (4th) 554,
[2003] 11 W.W.R. 244; Robinson v. Sydenham District Hospital Corp., [2000]
O.J. No. 703 (QL) at para. 30, 130 O.A.C. 109 (C.A.); Alderson v Callaghan
(1998), 40 O.R. (3d) 136 at p. 139, 111 O.A.C. 141, 42 C.C.L.T. (2d) 230 (C.A.);
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 225
Proving Causation where But For Unworkable 225
the damages will be reduced on that account, but there will be no
reverse onus requiring the plaintiff to prove that part or all of the injury
would have been sustained even had there been no contributory
negligence.44
The proviso involves the situation where another non-tortious cause
(or set of non-tortious causes)45 was a sufficient cause, meaning that
the same injury would have occurred in any event as a result of that
cause. That situation would not be one where the but for test is
unworkable. As the but for test in those circumstances would not be
satisfied the claim should be unrecoverable, in the same way that a
“crumbling skull” plaintiff — also a situation involving non-tortious
causes and where the but for test is not unworkable — is entitled to
recover only for that part of the ultimate injury that would not have
occurred in any event.
While some decisions have adopted that stance,46 others have held,
incorrectly in our view, either that a finding that the defendant’s wrongdoing is a materially contributing cause results in liability,47 or that a
determination must be made whether the defendant’s wrong-doing
constituted a materially contributing cause, thereby leading to liability.48
44.
45.
46.
47.
48.
Hosak v. Hirst, [2003] B.C.J. No. 107 (QL) at paras. 70-71, 291 W.A.C. 30, 9
B.C.L.R. (4th) 203 (C.A.); Larwill v. Lanham, [2003] B.C.J. No. 2627 (QL) at
para. 20, 311 W.A.C. 13, 21 C.C.L.T. (3d) 116 (C.A.), leave to appeal to S.C.C.
refused June 24, 2004; Kemp v. Wittenberg, [2001] B.C.J. No. 616 (QL) at paras.
49-50, 249 W.A.C. 13 (C.A.), supp. reasons 110 A.C.W.S. (3d) 33, application for
rehearing of cross-appeal refused 114 A.C.W.S. (3d) 791, leave to appeal to
S.C.C. refused [2003] 1 S.C.R. xii; T.W.N.A. v Clarke, [2003] B.C.J. No. 2747
(QL) at para. 22, 235 D.L.R. (4th) 13 sub nom. A. (T.W.N.) v. Clarke; M. (E.R.) v.
Clarke, [2004] 3 W.W.R. 11 (C.A.); Whitfield v. Calhoun, supra, footnote 26, at
para. 92; Santoro v. Raban, [2000] A.J. No. 494 (QL) at paras. 42-48, [2000] 7
W.W.R. 668, 79 Alta. L.R. (3d) 277 (Q.B.).
Heller v. Martens (2002), 213 D.L.R. (4th) 124 at paras. 44-47, [2002] 9 W.W.R.
71, 273 W.A.C. 84 (Alta. C.A.), likening the defendant’s argument to the last
clear chance doctrine. Cf. X v R.D.M., supra, footnote 40, at paras. 244-48, where
no reduction was made for a loss that was partly caused by the plaintiff’s illegal
conduct because the defendant’s wrongdoing was a materially contributing cause.
That is an incorrect decision in our view: see H.L. v. Canada (Attorney General),
supra, footnote 26, at paras. 137-43.
Where the other cause is a sufficient tortious cause, different considerations
apply. See the heading below.
Dushynski v. Rumsey, supra, footnote 39, at para. 17d; Aristorenas v. Comcare
Health Services, [2004] O.J. No. 3647 (QL) at para. 67, 133 A.C.W.S. (3d) 718
(S.C.). See also Sweeney v. O’Brien, [1990] N.S.J. No. 333 (QL), 99 N.S.R. (2d)
385 (S.C.), affd 31 A.C.W.S. (3d) 1298.
T.W.N.A. v. Clarke, supra, footnote 43, at para. 44.
B. (E.) v Order of the Oblates of Mary Immaculate, supra, footnote 24, at paras.
221 and 254; Chesher v. Monaghan, supra, footnote 26, at para. 56(3).
AQDavid(22)
7/22/2005
8:34 AM
Page 226
226 The Advocates’ Quarterly
[Vol. 30
6. The Situation in which the But For Test Is Truly,
although Only Partially, Abandoned
In a case where the same injury probably would have occurred in
any event due to the wrongdoing of one or more other persons, that
wrongdoing not being an unrelated intervening cause, application of
the but for test would have the anomalous result of excusing all of the
wrongdoers from liability, because each could legitimately take the
position that the injury would have occurred regardless of their fault.
The law will not countenance that type of unjust result. In such situations, the but for (or “even if”) test is truly abandoned, but only for the
purpose of rejecting the defence referred to above.49 It is not abandoned
insofar as the initial need to prove the element of causation with
respect to each of the individual wrongdoers is concerned.
7. Onus of Proof — the Normal Rule
As with all elements of a cause of action, the onus of proving
causation normally rests with the plaintiff.50 One explanation for this
is: “Historically, our legal system has placed the onus of proving
causation upon plaintiffs on the ground that defendants must not be
placed in the position of disproving claims asserted at the whim of
plaintiffs.”51
However, while the legal or ultimate burden of proof generally rests
with the plaintiff,52 “both the burden and the standard of proof are
flexible concepts”.53 “[C]ourts will strive to fashion a just solution . . .
to allow a wronged plaintiff to recover. Courts will not allow wronged
plaintiffs to fall between the cracks due to the formal requirements of
proving cause.”54
49. Fogg v. Kenora (Town), [1940] O.R. 421 at p. 423 (C.A.). See also Athey, supra,
footnote 3, at para. 22; Morris v. Collette, [2001] N.B.J. No. 298 (QL) (Q.B.) at
para. 164; Fairchild v. Glenhaven Funeral Services Ltd., supra, footnote 11, at
para. 10.
50. Athey, ibid., at para. 16; Walker Estate, supra, footnote 20, at
para. 87; Roncato, supra, footnote 35, at p. 722.
51. Gallant v. Fialkov, supra, footnote 30, at p. 317.
52. Snell, supra, footnote 12, at p. 301; Athey, supra, footnote 3, at para. 16; Brewer
Bros. v. Canada (Attorney General) (1991), 8 C.C.L.T. (2d) 45 at para. 79, 80
D.L.R. (4th) 321, [1992] 1 F.C. 25 (F.C.A.); Rhine v. Millan, [2000] A.J. No. 367
(QL) at para. 130, [2000] 7 W.W.R. 136, 78 Alta. L.R. (3d) 352 (Q.B.).
53. Snell, ibid., at p. 300.
54. Cottrelle v. Gerrard (2003), 67 O.R. (3d) 737 at para. 30, 233 D.L.R. (4th) 45,
178 O.A.C. 142 (C.A.).
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 227
Proving Causation where But For Unworkable 227
8. Reverse Onus of Proof
“Although the legal burden generally rests with the plaintiff, it is not
immutable . . . Valid policy reasons will be sufficient to reverse the
ordinary incidence of proof.”55
What has been described as “the classic illustration”56 of a situation
in which a reversal of the onus of proof is warranted are the shooting
incident cases. The Canadian example of such a case is Cook v.
Lewis,57 where the plaintiff was struck by a bullet fired by the gun of
one of two companions who shot simultaneously, and where the plaintiff was unable to show which of the two had fired the bullet. The
impossibility of proof led the court to find that both of the defendants
were liable. Sopinka J. provided the following rationale:58
Rather than letting the plaintiff fall between the two negligent defendants, the court reverses the onus of proof. Given the ability of the
defendants to crossclaim for contribution, this results in an equitable
splitting of the liability. Placing the burden of proof on the defendants
rather than the plaintiff was seen to be just. The reversal of onus was
justified as all possible causes of injury were before the court and had
been independently proven to be negligent in regard to this plaintiff. The
equities of the case therefore favoured the plaintiff over the negligent
defendants.
and the following warning:59
In such a case it is clear that the injury was not caused by neutral conduct.
It is quite a different matter to compensate a plaintiff by reversing
the burden of proof for an injury that may very well be due to factors
unconnected to the defendant and not the fault of anyone.
The reversal of the burden of proof is limited to the issue of linking
the damage to the defendant and does not apply to the issue of proving
a causal connection between the fault and the injury, and should only
be invoked where there is a true impossibility to determine the author
of the fault.60 That limitation, however, is inconsistent with the underlying justification for reversal of the onus of proof, as explained in the
following typical remark: “In our view, as a matter of policy, it would
be unjust to allow the [defendant] to escape liability by placing what,
55. Rainbow Industrial Caterers Ltd. v. Canadian National Railway Co., [1991] 3
S.C.R. 3 at para. 24, 84 D.L.R. (4th) 291, [1991] 6 W.W.R. 386.
56. Fairchild, supra, footnote 11, at para. 27.
57. [1951] S.C.R. 830.
58. “Whither Causation”, supra, footnote 9, at p. 8 (original emphasis).
59. Snell, supra, footnote 12, at p. 299.
60. Saint-Jean v. Mercier, [2002] 1 S.C.R. 491 at paras. 118 and 120, 209 D.L.R. (4th)
513 sub nom. St.-Jean v. Mercier (hereafter St-Jean).
AQDavid(22)
7/22/2005
8:34 AM
Page 228
228 The Advocates’ Quarterly
[Vol. 30
in some cases, could amount to an impossible burden on an innocent
plaintiff.”61
Thus, the burden of proof was effectively reversed in a case where
a manufacturer of breast implants argued that the plaintiff had the onus
of proving that, even had the former warned the plaintiff’s physician of
the dangers associated with the product, the physician would not then
have passed along that warning to the plaintiff. The court there said:
“Justice dictates that [the plaintiff] should not be penalized for the fact
that had the manufacturer actually met its duty to warn, the doctor still
might have been at fault.”62
In some cases, there is a shift of the evidentiary onus of proof. In a
claim for non-disclosure or negligent misrepresentation, once the
plaintiff has proved a loss arising from such wrongdoing, the onus then
falls upon the defendant to prove that the plaintiff would have sustained the same loss regardless of the breach.63 In a case where there is
a natural inference of causation from established facts, the defendant
must lead evidence to negative causation.64
9. Standard of Proof
Causation must be established on the normal civil standard of
balance of probabilities.65 It is not sufficient to demonstrate the mere
possibility of a causal connection,66 or to prove causation by way of
speculative evidence.67 The Ontario Court of Appeal has recently
reaffirmed this requirement.68
61. Walker Estate v. York Finch General Hospital (1999), 43 O.R. (3d) 461 at p. 474,
169 D.L.R. (4th) 689, 118 O.A.C. 217 (C.A.), affd supra, footnote 20.
62. Hollis v. Dow Corning Corp. (1995), 129 D.L.R. (4th) 609 at paras. 53-57, [1995]
4 S.C.R. 634, [1996] 2 W.W.R. 77 sub nom. Hollis v. Birch.
63. Hodgkinson v. Simms, supra, footnote 18, at p. 200; Canada Trustco Mortgage
Co. v. Bartlet & Richardes (1996), 28 O.R. (3d) 768 at p. 775, 91 O.A.C. 33
(C.A.). Cf. Canada Trust Co. v. Sorkos (1992), 90 D.L.R. (4th) 265 (Gen. Div.) at
p. 272, affd 135 D.L.R. (4th) 383, 3 R.P.R. (3d) 262 (Ont. C.A.).
64. Stewart v. Pettie, supra, footnote 7, at para. 68. See also Branscombe v. Regina
District Health Board, [2003] S.J. No. 541 (QL) at paras. 29-30, [2004] 3 W.W.R.
624, 305 W.A.C. 59 sub nom. Branscombe v. Greeff.
65. Athey, supra, footnote 3, at para. 13; RJR-MacDonald Inc. v. Canada (Attorney
General) (1995), 127 D.L.R. (4th) 1 at para. 137, [1995] 3 S.C.R. 199, 100 C.C.C.
(3d) 449; Allen (Next Friend of), supra, footnote 6, at para. 38; Loder, supra,
footnote 8, at para. 59; Cameron (Litigation Guardian of), supra, footnote 8, at
para. 159.
66. Rothwell v. Raes (1990), 2 O.R. (3d) 332 at p. 335, 76 D.L.R. (4th) 280 (C.A.),
leave to appeal to S.C.C. refused 79 D.L.R. (4th) vii; Johnston v. Murchison,
[1995] P.E.I.J. No. 23 (QL) at para. 63, 127 Nfld. & P.E.I.R. 1 (C.A.).
67. Meloche v. Hotel Dieu Grace Hospital/Villa Marie, supra, footnote 32, at para. 38.
68. Cottrelle v. Gerrard , supra, footnote 54, at paras. 23-26 and 35.
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 229
Proving Causation where But For Unworkable 229
An exception is made, however, for hypothetical and future events
that cannot and therefore need not be proved on a balance of probabilities. Instead, they are given weight in accordance with their relative
likelihood, provided they are real and substantial possibilities and not
mere speculation.69 This exception usually arises in the context of an
assessment of damages,70 particularly in “crumbling skull” cases.71
The exception has also been applied in the context of proof of
causation,72 although the usual approach is that the issue of causation
involves past events and factual rather than hypothetical matters, so that
the normal standard of proof is applicable.73 That approach is reflected
in the rule that loss of a chance — i.e. less than a 50% likelihood — is
not sufficient to establish causation,74 although this may be contrasted
with authorities regarding loss of opportunity claims.75 While it is
arguable that the latter are distinguishable in that they involve the issue
of assessment of damages rather than causation, the line between the two
is exceedingly fine and does not warrant the difference in treatment.
Although it has been said that the standard of proof for causation is
a flexible concept,76 that statement apparently was not intended to
mean that a standard lower than that of balance of probabilities can be
applied, but rather that the manner of satisfying that standard of proof
should be expanded.
69. Athey, supra, footnote 3, at para. 27.
70. Fraser Park South Estates Ltd. v. Lang Michener Lawrence & Shaw, [2001]
B.C.J. No. 21 (QL) at para. 80, 239 W.A.C. 86, 84 B.C.L.R. (3d) 65 (C.A.), leave
to appeal to S.C.C. refused 263 W.A.C. 320n and Wolfstadt v. Chaiton & Chaiton
(2003), 65 O.R. (3d) 531 at paras. 37-38 (S.C.), where different standards were
applied.
71. Such as Lyne v. McClarty, [2003] M.J. No. 29 (QL) at para. 27, [2003] 5 W.W.R.
598, 285 W.A.C. 161 (C.A.); Dufty v. Great Pacific Industries Inc., [2000] B.C.J.
No. 1988 at para. 121, 99 A.C.W.S. (3d) 1083 (S.C.).
72. Kelly v. Lundgard, [2001] A.J. No. 906 (QL) at para. 25, 202 D.L.R. (4th) 385,
[2001] 9 W.W.R. 399 (C.A.).
73. Sales v. Clarke, [1998] B.C.J. No. 2334 (QL) at paras. 10-16, 165 D.L.R. (4th)
241, 184 W.A.C. 26 (C.A.); Kemp v. Wittenberg, supra, footnote 43, at paras. 4248 (C.A.); St-Jean, supra, footnote 60, at para. 106.
74. St-Jean, ibid., at para. 106; Cottrelle, supra, footnote 54, at para. 36; Fournier v.
Wiens, supra, footnote 26, at para. 98; Lindahl Estate v. Olsen, supra, footnote 35,
at para. 33; “Whither Causation”, supra, footnote 9, at p. 6.
75. Eastwalsh Homes Ltd. v. Anatal Developments Ltd. (1993), 12 O.R. (3d) 675 at
pp. 687-90, 100 D.L.R. (4th) 469, 62 O.A.C. 20 (C.A.), leave to appeal to S.C.C.
refused 104 D.L.R. (4th) vii; Smith v. Knudsen, [2004] B.C.J. No. 2509 (QL) at
paras. 22-38, 247 D.L.R. (4th) 256, [2005] 1 W.W.R. 216 (C.A.); Rodaro v. Royal
Bank of Canada (2002), 59 O.R. (3d) 74 at paras. 57ff, 157 O.A.C. 203, 22 B.L.R.
(3d) 274 sub nom. Rodaro v. Royal Bank (C.A.); Lindahl Estate, ibid., at paras. 34,
146-149 and 430.
76. Snell, supra, footnote 12, at p. 300.
AQDavid(22)
7/22/2005
8:34 AM
Page 230
230 The Advocates’ Quarterly
[Vol. 30
There are, however, decisions that suggest that the standard of proof
itself is flexible. In one Supreme Court of Canada decision involving
injury to a student who was operating a power saw without supervision,
the following statement was made: “I do not find it improbable that the
accident would not have happened if the instructor had directly supervised the operations until they were finished.”77 That curiously phrased
formulation, couched in negative language, suggests that the normal
standard would not have been satisfied on the evidence.
A later decision of that court, also involving an injury to an
unsupervised student (this time in gym class) adopted that statement,
but in the course of its reasons the court went considerably further in
relaxing the required standard of proof:78
In my opinion, it cannot be said that the presence of a teacher among six
to eight students in the exercise room would not have had a restraining
effect upon the students which could have influenced the course of events
and prevented the accident . . . In my opinion, it would be more than mere
speculation to conclude, as did the trial judge, that the presence of a
teacher in the exercise room could have influenced the situation.
The standard applied in that case appears to involve the real and substantial possibility standard that is applied to proof of hypothetical and
future events.
10. The Means by which Causation Can Be Proved
How then does one establish that a person’s negligence “materially
contributed” to the occurrence of an injury when the but for test is
unworkable?
Causation is “essentially a practical question of fact which can best
be answered by ordinary common sense”.79 Not only is scientific precision not required, neither is scientific proof.80 If opinion evidence is
led, a trial judge may well be influenced by it but is not bound to accept
all or any of it81 including, for example, evidence that there was only a
30% statistical likelihood of a causal connection.82 The causation test
77. Dziwenka v. Alberta, [1972] S.C.R. 419 at p. 433, 25 D.L.R. (3d) 12, [1972] 1
W.W.R. 350.
78. Myers v. Peel County Board of Education, supra, footnote 39, at p. 13 (emphasis
added).
79. Snell, supra, footnote 12, at p. 300; Athey, supra, footnote 3, at para. 16; Hock,
supra, footnote 12, at para. 120.
80. Snell, ibid., at pp. 301 and 306; Athey, ibid., at para. 16; Hock, ibid., at para. 120.
81. Canada Post Corp. v. Nova Scotia (Workers’ Compensation Appeals Tribunal),
[2004] N.S.J. No. 242 (QL) at para. 24 (C.A.); Laferriere v. Lawson (1991), 78
D.L.R. (4th) 609 at p. 657, [1991] 1 S.C.R. 541, 38 Q.A.C. 161.
82. “Whither Causation”, supra, footnote 9, at p. 2. On the other hand, in Lurtz v
Duchesne, supra, footnote 13, at paras. 370-72, vard Ont. C.A., February 4, 2005,
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 231
Proving Causation where But For Unworkable 231
is not to be applied too rigidly.83 In cases where it is difficult to prove
causation, a “robust and pragmatic approach” should be taken.84
While the purpose of this approach is to avoid plaintiffs being left
uncompensated for technical reasons, it does not mean that the law is
“subject to judicial caprice” but rather that “the cases are very factdriven”.85 “The conclusion must be rooted in the evidence and not the
product of mere speculation.”86
Where the parties have adduced expert evidence on the issue of
causation,87 or where the evidence affirmatively demonstrates an
absence of causation,88 it is not open to the court to apply the “robust
and pragmatic approach”.
In cases where there is a practical impossibility to prove causation
by normal types of evidence, the most common alternative means of
proof is by way of inference. Snell gave a stamp of approval for various
routes through which proof by way of inference could be made.
An inference of causation can be drawn as a matter of common
sense.89 This is often done in cases involving medical malpractice
claims.90 Other examples include cases in which the negligence of
Transport Canada was causally connected to an airplane crash91 and the
failure to stop home delivery of a newspaper to a break-in.92 A close
temporal connection between the wrongdoing and the injury is a
factor in drawing an inference of causation on the basis of common
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
evidence of an 80% likelihood that, had surgery on an undiagnosed ailment been
performed it would have been successful, established causation.
Snell, supra, footnote 12, at p. 301; Athey, supra, footnote 3, at para. 16.
B.M., supra, footnote 10, at para. 127; Montgomery v. Luoma, [2001] B.C.J.
No. 1950 (QL) at para. 38, 256 W.A.C. 267, 92 B.C.L.R. (3d) 195 (C.A.);
Meloche, supra, footnote 32, at para. 21; Aristorenas v. Comcare Health Services,
supra, footnote 46, at para. 68.
Phillip v. Whitecourt General Hospital, supra, footnote 21, at paras. 242-43
(Q.B.).
Cameron (Litigation Guardian of), supra, footnote 8, at para. 161.
Miller v. Budzinski, supra, footnote 31, at para. 474. This ought not to apply if the
judge rejects all of the expert evidence.
Cottrelle, supra, footnote 54, at para. 33; Lynne v. Taylor, [2004] A.J. No. 755
(QL) at para. 150, 132 A.C.W.S. (3d) 691 (Q.B.); Oiom v. Brassington, [1990]
B.C.J. No. 2686 (QL), 52 B.C.L.R. (2d) 240 (C.A.), leave to appeal to S.C.C.
refused 135 N.R. 201n.
RJR-MacDonald, supra, footnote 65, at para. 137.
Such as Robinson v. Sydenham District Hospital Corp., supra, footnote 43, at
para. 27; Lurtz v. Duchesne, Ont. C.A., supra, footnote 13, at para. 12;
Aristorenas, supra, footnote 46, at para. 73.
Swanson Estate v. Canada, supra, footnote 32, at pp. 758-59.
Crotin v National Post, Southam Publications (2003), 67 O.R. (3d) 752, 20
C.C.L.T. (3d) 316 (Small Claims Ct.).
AQDavid(22)
7/22/2005
8:34 AM
Page 232
232 The Advocates’ Quarterly
[Vol. 30
sense, although not a conclusive factor,93 and conversely a lengthy time
period94 between the two is a factor in drawing the common sense inference of an absence of causal connection.95 On the other hand, common
sense may lead away from making an inference of causation.96
A slightly different formulation of the common sense approach is
that an inference of causation may be drawn on the basis of reason or
logic. “The existence of scientific proof is simply of probative value in
demonstrating this reason, logic or common sense. It is by no means
dispositive or determinative.”97
An inference may be drawn for other reasons as well. The defendant’s wrongdoing may have rendered it impossible for the plaintiff
to prove causation.98 An absence of evidence from the defendant to
rebut evidence led by the plaintiff that has created a “presumptive
causal link”99 is a factor in the drawing of an inference of causation,100 particularly where the defendant is uniquely able to lead such
evidence.101 The materialization of a clear danger may lead to an
inference of causation, unless there is a demonstration or indication
to the contrary.102
93. As in Sparkes-Morgan v. Webb, [2002] N.J. No. 241 (QL) at para. 14, 216 Nfld.
& P.E.I.R. 112, 28 M.V.R. (4th) 233 (C.A.); Chriqui v. Duke, [2004] A.J. No. 697
(QL) at paras. 64 and 80, 31 Alta. L.R. (4th) 324 (Q.B.).
94. Lengthy in relation to the circumstances.
95. Anderson v. McAndrew, [2003] A.J. No. 24 at para. 65 (QL), [2003] 5 W.W.R.
249, 9 Alta. L.R. (4th) 143 (Q.B.); Loder, supra, footnote 8, at paras. 61-64.
96. As in Hammond v. Wabana (Town), [1998] N.J. No. 336 (QL) at para. 50, 44
N.S.R. (2d) 274 (C.A.); Sherman v. Salsberg, [1998] O.J. No. 3074 (QL) at paras.
105-106, 81 A.C.W.S. (3d) 730 (Gen. Div.).
97. RJR-MacDonald, supra, footnote 65, at paras. 153-57 and 184.
98. St-Jean, supra, footnote 60, at para. 118; Snell, supra, footnote 12, at p. 305;
Levitt v. Carr, [1992] B.C.J. No. 704 (QL), [1992] 4 W.W.R. 160, 23 W.A.C. 27
(C.A.), leave to appeal to S.C.C. refused 28 W.A.C. 82n, where the issue was
framed in terms of a material increase in the risk. In Bigcharles v. Dawson Creek
and District Health Care Society, [2001] B.C.J. No. 1003 (QL) at paras. 61-71,
251 W.A.C. 134, 91 B.C.L.R. (3d) 82 (C.A.), leave to appeal to S.C.C. refused
[2002] 1 S.C.R. v sub nom. Bigcharles v. Lomax, it was held that the trial judge
was entitled on the evidence not to draw an inference on this basis.
99. Walker Estate v. York-Finch General Hospital, supra, footnote 20, at paras. 8586. Conversely, the evidence may not have established that “presumptive link”:
Robb v. St. Joseph’s Health Centre; Rintoul v. St. Joseph’s Health Centre; Farrow v.
Canadian Red Cross Society, [2001] O.J. No. 4605 (QL) at para. 69 (C.A.).
100. Robinson, supra, footnote 43, at para. 25.
101. Snell, supra, footnote 12, at p. 300; Rhine, supra, footnote 52, at para. 130. The
defendant may not have a monopolistic control over the relevant evidence, as in
Robb v. St. Joseph’s Health Centre, supra, footnote 99, at para. 71.
102. Laferriere v. Lawson, supra, footnote 81, at pp. 658-59.
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 233
Proving Causation where But For Unworkable 233
An inference of causation will not be drawn merely because it is
practically impossible for the plaintiff to prove causation.103 Nor will
an inference be drawn if an equally plausible theory is presented in the
evidence which does not involve a causal connection.104 Furthermore,
when seeking to draw an inference, the following admonition should
be heeded: “There is a fine line between inference and speculation. The
court must be careful to draw only those inferences which are supported
by objective evidence.”105
11. Justification for Relaxation of the Means of Proof
The justification for relaxation of any rule of law is, of course, the
desire to achieve a fair and just result in the circumstances of the case.
Among the many statements that have been made in that vein in the
context of causation are the following:
[C]ourts will strive to fashion a just solution in this type of case to allow
a wronged plaintiff to recover. Courts will not allow wronged plaintiffs to
fall between the cracks due to the formal requirements of proving
cause.106
Indeed, it would seem to me contrary to principle to insist on application
of a rule which appeared, if it did, to yield unfair results.107
[I]f the defendants’ own wrongdoing prevents the plaintiff from making
the necessary causal connection to the specific author of the wrong, liability is to be attributed collectively so as to avoid the injustice of leaving the victim with no recourse.108
The legal or ultimate burden of proof is determined by the substantive
law “upon broad reasons of experience and fairness”.109
12. Is Further Relaxation Warranted?
In a well-known passage in Snell, Sopinka J. said the following: “If
I were convinced that defendants who have a substantial connection to
the injury were escaping liability because plaintiffs cannot prove
causation under currently applied principles, I would not hesitate to
103. Dixon Estate v. Holland, [1996] B.C.J. No. 1252 (QL) at para. 22, [1997] 1
W.W.R. 130, 25 B.C.L.R. (3d) 278 (S.C.).
104. Dixon Estate, ibid., at para. 21.
105. Johnston v. Murchison, supra, footnote 66, at para. 74 . “Post hoc ergo proper hoc
has no place in our law.” Davy v. Morrison, [1931] 4 D.L.R. 619 (Ont. C.A.) at p.
626.
106. Cottrelle, supra, footnote 54, at para. 30, quoting from Lewis N. Klar, Tort Law,
3rd ed. (Toronto, Thomson Carswell, 2003).
107. Fairchild, supra, footnote 11, at para. 13.
108. St-Jean, supra, footnote 60, at para. 118.
109. Snell, supra, footnote 12, at p. 294, quoting from J.H. Wigmore, Evidence in
Trials at Common Law, 4th ed. (Boston, Little Brown & Co., 1981).
9 — 30 A.Q.
AQDavid(22)
7/22/2005
8:34 AM
Page 234
234 The Advocates’ Quarterly
[Vol. 30
adopt one of [two specified] alternatives.”110 In his later writing, he
referred to that statement and expanded on it, saying: “That, my
friends, is not a threat, but rather a not-too-subtle hint as to ‘Whither
Causation’.”111
McLachlin C.J.C. made a similar comment in a paper published in
1998, which was quoted in the Fairchild decision: “[T]oo often the
traditional ‘but for’, all-or-nothing, test denies recovery where our
instinctive sense of justice — of what is the right result for the situation — tells us the victim should obtain some compensation.”112
Has the time come, not, as McLachlin C.J.C. implied, to abandon
the but for test in the type of situation described by her, but to further
relax the rules regarding the manner in which that test, and through it
the element of causation, can be proved? Are there new approaches
that should be adopted to deal with the particularly difficult situations
(and only such situations) that would establish what has been described
as a “presumptive causal link”113 that would then generate an evidentiary reverse onus of proof, requiring the defendant to show that in fact
there was no causal connection between the wrongdoing and the
injury?
Before addressing that question, reference should be made to the
danger of over-relaxation of the rules. Lord Nicholls made the following
statement in Fairchild:114
I need hardly add that considerable restraint is called for in any relaxation
of the threshold “but for” test of causal connection. The principle applied
on these appeals is emphatically not intended to lead to such a relaxation
whenever a plaintiff has difficulty, perhaps understandable difficulty, in
discharging the burden of proof resting on him. Unless closely confined
in its application this principle could become a source of injustice to
defendants. There must be good reason for departing from the normal
threshold “but for” test. The reason must be sufficiently weighty to justify
depriving the defendant of the protection this test normally and rightly
affords him, and it must be plain and obvious that this is so. Policy questions will loom large when a court has to decide whether the difficulties
of proof confronting the plaintiff justify taking this exceptional course.
Again, it is our view that further relaxation would not involve a “departure” from the but for test, but rather a moderation of the manner by
110. Snell, ibid., at p. 299.
111. “Whither Causation”, supra, footnote 9, at pp. 11-12.
112. Fairchild, supra, footnote 11, at para. 11.
113. Walker Estate v. York-Finch General Hospital, supra, footnote 20, at paras. 85-86.
114. Fairchild, supra, footnote 11, at para. 43. See also paras. 33 and 169. In Snell,
supra, footnote 12, reference was made at p. 299 to the contribution made by liberalization of the rules in the United States to the medical malpractice insurance
crisis there in the 1970s.
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 235
Proving Causation where But For Unworkable 235
which satisfaction of that test can be presumptively proved. Some
might argue, with a measure of justification, that the difference is more
illusory than real, and that further relaxation would have the effect of
abandonment of the but for test. Our response to that argument is that
retention of the but for test, even if only in theory and semantically,
is required to maintain the foundational principle115 that a person is
liable only for those injuries that he or she has caused or materially
contributed to. One need only be reminded of the “first principles”
comment made in Athey to understand the critical importance of the
but for test in all tort claims.116
A “presumptive” causal connection is nonetheless a causal connection. Furthermore, the very purpose of further relaxation is to deal with
the kinds of situation referred to by Sopinka J. and McLachlin C.J.C.,
where there is a genuine belief that a causal link exists, but that link is
practically unprovable, even by the already-relaxed rules as established in Athey. Finally, the same criticism might well be directed at
the manner in which the Snell approach may be applied, where an
inference might be drawn, because of the genuine but unprovable
belief that a causal connection exists, in circumstances that do not truly
warrant doing so.117
The situation in Fairchild may well be an example of a situation
warranting a further relaxation of the rules. The factual background
there118 was a variation of the Cook v. Lewis scenario, the added complication being that the plaintiff was unable, because of the current
limits of human science, to show that his injury was necessarily caused
by one or the other of the defendants, or even by a combination of the
wrongdoing of both of them. Perhaps resort should have been had to
the common sense inference of causation approach, which presumably
would have led to a finding that the injury would not have occurred but
for the combined negligence of the two defendants, thereby squarely
making it a Cook v. Lewis situation and placing the onus of proof
on the defendants to show that their wrongdoing did not cause or
materially contribute to the injury. A contributing cause need not be a
sufficient cause,119 so that the inability to prove that the wrongdoing of
each individual defendant was not in itself a sufficient cause of the
injury would have been irrelevant. The issue of relative causation
115. As it was described by Sopinka J. See footnote 9.
116. See supra, footnote 3.
117. As in Scott v. Mohan, [1993] A.J. No. 592 (QL) at paras. 120-34 (Q.B.).
118. The factual background of Fairchild is summarized at para. 2 of the decision,
supra, footnote 11.
119. See footnote 39.
AQDavid(22)
7/22/2005
8:34 AM
Page 236
236 The Advocates’ Quarterly
[Vol. 30
would have been a matter of apportionment between the defendants.
The House of Lords chose, however, not to take that route but rather
resurrected the relaxation of proof approach adopted in their wellknown earlier decision in McGhee v. National Coal Board,120 which
subsequently had been disapproved in their similarly well-known
decision in Wilsher v. Essex Area Health Authority.121
Whether or not Fairchild is an appropriate example of the kind of
situation that calls for a further relaxation of the rules, it is virtually
inevitable that that kind of case will arise, perhaps more frequently
than might presently be imaginable, as implied in the comments of
McLachlin C.J.C.122 The area of medical malpractice in particular is
rife with possibilities for that kind of situation, and the McGhee case
was itself an example of it.123
13. Examples of More Relaxed Rules
Should there be a situation that calls for more relaxed rules, a choice
might be made from among the following:
(a) The McGhee test: did the injury occur within the area of a risk created or materially increased by the defendant’s wrongdoing?
While this test has been criticized as effectively eliminating the
need to prove causation124 and was apparently, but not expressly,
rejected in Snell (although it was more clearly rejected in a subsequent decision of the Supreme Court of Canada),125 its resurrection
in Fairchild may lead to second thoughts in Canada as well. There
are, in fact, several cases since Snell where it has either been
referred to or applied,126 and there is little to distinguish it from the
“materialization of a clear danger” approach taken in another
Supreme Court of Canada decision.127
120. [1973] 1 W.L.R. 1 (H.L.).
121. [1988] A.C. 1074 (H.L.).
122. See supra, footnote 112.
123. Fairchild, supra, footnote 11, at para. 21.
124. B.M. v British Columbia (Attorney General), supra, footnote 10, at para. 156,
quoting from Klar, Tort Law, supra, footnote 106.
125. St-Jean, supra, footnote 60, at para. 116. See also Allen (Next Friend of) v.
University Hospitals Board, supra, footnote 6, at para. 19.
126. Robinson v. Sydenham District Hospital Corp., supra, footnote 43, at para. 20;
Cameron (Litigation Guardian of) v. Louden, supra, footnote 8, at paras. 155,
268, 315, 366, 382 and 385; Anderson v. McAndrew, supra, footnote 95, at
para. 65; Lindahl Estate v. Olsen, supra, footnote 35, at paras. 120-23; Rhine v.
Millan, supra, footnote 52, at paras. 132 and 185 (Q.B.).
127. See supra, footnote 102.
AQDavid(22)
7/22/2005
2005]
8:34 AM
Page 237
Proving Causation where But For Unworkable 237
(b) The second part of the Amos test:128 is there some nexus or causal
relationship between the wrongdoing and the injury, or is the
connection merely incidental or fortuitous? This test was established to deal with statutory language involving injuries which
“arise out of the use or operation of a vehicle”.
(c) The “rational connection” test: was the injury rationally connected
to the wrongdoing? This is a test employed in certain constitutional
law cases.129
(d) The “proximity analysis” test: a case-by-case analysis employed in
other constitutional cases that measures the entire relationship
between the breach and the injury.130
(e) The “substantial connection to the injury” test referred to in Snell
itself.131
There is no reason why any one of these tests cannot and should not
be adopted in the appropriate tort situation where the but for test is
unworkable and where justice so requires, to prove on a presumptive
basis the necessary “but for” causal connection.
14. Conclusion
Causation is a necessary element in every tort cause of action. The
“but for” concept is, in turn, an essential ingredient of causation and
therefore must be satisfied in all cases. In some situations, however,
that proof is practically impossible. The response to such situations is
not an inquiry whether the defendant’s wrongdoing constituted a
materially contributing cause of the injury but rather a relaxation of the
means for satisfaction of the but for test. Proof can be made, for
example, by way of an inference drawn through common sense.
There will, however, be instances where even the relaxed means of
proof as presently permitted will not be sufficient to demonstrate satisfaction of the but for test. In the limited category of situations where,
to use the language of Snell, “defendants who have a substantial
128. Amos v. Insurance Corp. of British Columbia, [1995] 3 S.C.R. 405 at p. 415, 127
D.L.R. (4th) 618, [1995] 9 W.W.R. 306.
129. See, for example, RJR-MacDonald Inc. v. Canada (Attorney General) supra,
footnote 65, at paras. 82ff and 153ff.
130. See, for example, R. v. Goldhart, [1996] 2 S.C.R. 463 at paras. 34-40, 136 D.L.R.
(4th) 502, 107 C.C.C. (3d) 481.
131. Snell, supra, footnote 12, at p. 299.
AQDavid(22)
7/22/2005
8:34 AM
Page 238
238 The Advocates’ Quarterly
[Vol. 30
connection to the injury would be escaping liability because plaintiffs
cannot prove causation under currently applied principles”,132 there
should be a further relaxation of the means by which satisfaction of the
but for test, and thereby proof of causation, can be made. Some
examples of such further relaxation are listed above. Utilization of those
more relaxed means would result in a presumptive causal connection
that would generate an evidentiary reverse onus of proof, requiring the
defendant to show that in fact there was no causal connection.
132. See supra, footnote 110.